Community Violence Concerns and Adolescent Sleep: Physiological Regulation and Race as Moderators

Abstract

Prior work has demonstrated that greater community violence concerns are associated with poor sleep quality among adolescents. However, these effects may not be uniform across all youth. The present study examined the role of individual difference variables, physiological regulation and race, as moderators of risk in the relation between adolescents’ community violence concerns and their sleep. Adolescents (N =219; 55.3% female; 69.9% White/European American, 30.1% Black/African American) participated in the study when they were 18 years old (M = 17.7 years, SD = 1.0). Physiological regulation was assessed via respiratory sinus arrhythmia, a measure of parasympathetic regulation, at rest and in response to a stressor. Adolescents wore actigraphs for 7 nights to assess their sleep duration and quality and reported on their community violence concerns via a well-validated questionnaire. Results demonstrated a consistent pattern of interactions, such that African American adolescents who showed less adaptive patterns of regulating physiological arousal experienced shorter sleep duration and poorer sleep quality in the context of greater community violence concerns. Community violence concerns were not associated with sleep for White adolescents. The findings may suggest that race-related stressors exacerbate risk for poor sleep among African American adolescents who experience more community violence concerns and have more difficulty regulating physiological arousal. Coping strategies for managing stress and arousal may be helpful for improving sleep for some youth.

INTRODUCTION

The majority of American adolescents ages 14–17 report witnessing violence in their lifetime (71.5%; Finkelhor et al., 2015). Exposure to such violence may elicit concerns for safety that elevate arousal levels, which can make it difficult to initiate and maintain sleep (Heissel et al., 2018). However, other variables may also play a role in these associations. For example, individual differences in the ability to regulate physiological arousal are linked to duration and quality of sleep (Elmore-Staton et al., 2012; El-Sheikh et al., 2013). Furthermore, in comparison to White/European American (EA) adolescents, Black/African American (AA) adolescents are at greater risk for shorter and poorer quality sleep (Moore et al., 2011). The present study examined the extent to which the association between community violence concerns and risk for short and poor quality sleep, as measured through actigraphy, differed depending on adolescents’ physiological regulation and race.

A small body of literature has detected associations between community violence concerns and adolescent sleep disruption. Using a sample of 16-year-olds, Bagley and colleagues (2016) found that greater community violence concerns were linked to subjective and actigraph-assessed poor sleep quality. Another study with 14-year-olds found neighborhood distress, calculated via indices including higher poverty and high school dropout rates, to be associated with actigraphy-based fewer sleep minutes and greater variability in sleep schedule (Moore et al., 2011). These relations were attenuated when controlling for race. Worries about community violence may lead to increases in physiological arousal that are incompatible with sleep (Heissel et al., 2018).

Adolescents who are exposed to community violence may experience physiological hyperarousal (Fowler et al., 2009). Elevated nighttime blood pressure has been detected in AA adolescents who reported victimization experiences, and boys who reported hearing of community violence had high levels of daytime epinephrine (Wilson et al., 2002). Greater ability to regulate physiological arousal may protect against poor sleep in the context of stress associated with community violence concerns. Physiological regulation can be assessed via respiratory sinus arrhythmia (RSA), a measure of parasympathetic nervous system (PNS) activity. According to Porges’ Polyvagal Theory, higher baseline RSA corresponds to greater parasympathetic relative to sympathetic control of the heart, and underlies flexible responding to environmental stimuli (Porges, 2007). A greater decrease in RSA in response to challenge (i.e., suppression) indicates a reduction of parasympathetic influence on cardiac activity that may allow for mobilization of sympathetic resources that underlie adaptive cognitive, emotional, and behavioral responses to the stressor.

Higher baseline RSA is associated with higher nighttime sleep efficiency among preschool-aged children (Elmore-Staton et al., 2012) and with higher self-reported sleep quality among adult men (Irwin et al., 2006). Greater RSA suppression is also linked to fewer subjective ratings of sleep problems in infancy (Gueron-Sela et al., 2017) and better subjective and actigraph-assessed sleep quality in middle childhood (El-Sheikh et al., 2013). Furthermore, children who are better able to regulate physiological arousal may sleep better in contexts of stress. For example, 9–10 year-old children with poor initial sleep who exhibited greater RSA suppression in response to challenge and who lived with families experiencing high marital conflict showed improvement in sleep across one year (El-Sheikh et al., 2015).

In addition to physiological regulation, an individual’s race/ethnicity may influence the association between community violence concerns and sleep. AA adolescents have shorter and poorer quality sleep relative to EAs (Moore et al., 2011). As people of color living in the United States, AAs experience race-based stressors such as discrimination that increase vulnerability to poor sleep (Tomfohr et al., 2012). Consistent with cumulative risk and health disparities perspectives (Gee et al., 2012), race-based stress may contribute additional risk that increases AA adolescents’ vulnerability to poor sleep when combined with difficulty managing physiological arousal in the context of community violence concerns.

The present study extends prior work by examining physiological regulation and race as moderators of the association between community violence concerns and adolescents’ sleep. Towards a thorough investigation of effects, multiple aspects of physiological regulation (baseline, response to challenge) and actigraphy-based sleep (duration, efficiency) were examined. Our hypotheses were as follows:

1. Lower baseline RSA or less RSA suppression were expected to increase risk for poor sleep in the context of greater community violence concerns.

2. AA adolescents were expected to be more vulnerable to poor sleep than EA adolescents in the context of greater community violence concerns.

3. AA adolescents who had lower baseline RSA or showed less RSA suppression to challenge and who also reported greater community violence concerns were hypothesized to be at greatest risk for poor sleep.

METHODS

Participants

Data were collected as part of the Family Stress Study at Auburn University. Neither actigraphy nor community violence data were available at the beginning of the larger longitudinal investigation. Participants were recruited from local school districts in 2005 and were included in the study on the conditions of living in a two-parent home and not having diagnoses of attention-deficit/hyperactivity disorder, developmental delays, chronic illnesses, or sleep disorders. The 219 (M_age = 17.7 years, SD = 1.0) adolescents in the sample were representative of the recruitment area (U.S. Census Bureau, 2005) and varied with respect to sex (55.3% female), race (69.9% White/European American, 30.1% Black/African American), and socioeconomic status (36.1% living at or below the federal poverty line, 22.8% lower middle class, and 41.1% middle class). Socioeconomic status (SES) was measured by income-to-needs ratio, the quotient of a family’s total yearly income divided by the poverty threshold for the size of their household as determined by the U.S. Census Bureau (2018).

Procedures

The university’s institutional review board approved study procedures. Parents gave written consent, and adolescents assented to participate. Adolescents wore actigraphs at home for one week, then completed a physiological assessment and questionnaires in a campus laboratory. The majority (66.1%) of participants visited the laboratory the day following the last night of wearing the actigraph, and an average of .34 days (SD = 5.37 days) separated the home actigraph-wear and laboratory assessment. During the lab visit, trained researchers placed electrodes on the adolescent following established procedures (Hinnant et al., 2015). After a 3min acclimation, a 3min RSA baseline was obtained, followed by RSA regulation during a 3min challenge task. Questionnaires were then completed on a computer.

Measures

Community violence concerns

Adolescents completed the community violence scale of the well-established and validated Community Experiences Questionnaire (Schwartz & Proctor, 2000). Participants responded to 7 items regarding their community, such as “How worried are you that someone will break in or force their way into your home?”; “How worried are you that someone will hurt you really badly?”; and “How worried are you that someone will steal something from you using violence?” on a scale from 0 = “Not at all” to 4 = “A whole lot.” Summed scores were used in analyses. Reliability for the scale was high (α = .90).

Respiratory sinus arrhythmia

Two well-established indices of RSA were assessed: a baseline level (RSAB) and a regulation score (RSAR). RSA data were collected and analyzed following well-validated procedures (Hinnant et al., 2015). Cardiovascular data were collected using a Mindware BioNex 8-slot chassis and MW1000A acquisition system (Mindware Technologies Ltd, Gahanna, OH, USA) connected to an electrocardiograph (ECG) activity amplifier module and disposable pediatric snap ECG electrodes, and respiration was derived from spectral analysis of thoracic impedance (Z₀; Ernst et al., 1999). Data were scored in 1min intervals using Mindware analysis software (HRV version 3.0.22). A trained researcher reviewed the cardiovascular data and manually edited artifacts and missing or misplaced R-peaks. Basal RSA was derived from the natural log of the high-frequency power (0.15–0.40 Hz), a validated method used to isolate parasympathetic influence on cardiac activity (Berntson et al., 1997). Basal RSA (RSAB) was obtained during 3min of rest following acclimation. RSAR was measured during the Iowa Gambling Task (IGT), a stressor which has been shown to elicit physiological arousal (Bechara et al., 1997), in which participants took risks while selecting from four decks of cards in an effort to earn points toward a monetary reward (a $10 gift card). RSA level during the task was subtracted from the basal level to calculate RSAR; a lower (negative) score represents a greater decrease in RSA (i.e., greater suppression) and parasympathetic influence and a more optimal response. Both RSAB and RSAR have been shown to be moderately to highly stable across adolescence (Hinnant et al., 2018).

Race

Adolescents’ race was reported by parents at the time of data collection and coded as 0 = White/European American and 1 = Black/African American.

Sleep

Adolescents wore Motionlogger Octagonal Basic actigraphs (Ambulatory Monitoring Inc., Ardsley, NY, USA) on their non-dominant wrists while sleeping for up to 7 consecutive nights during the school year. Data were obtained in 1min epochs using zero-crossing mode and scored using the Sadeh algorithm (Sadeh et al., 1995) to derive two sleep parameters: minutes and efficiency. Sleep minutes signify the number of minutes scored as asleep between sleep onset and wake. Sleep efficiency, represented as a percentage, is derived by dividing sleep minutes by the total number of minutes between sleep onset and offset. For each parameter, the mean across all nights was used in analyses. On average, participants had 5.56 nights of sleep data (SD = 1.28). The most common reason for missing nights of data was participants forgetting to wear the actigraph; additionally, researchers excluded nights during which adolescents took medication for an acute illness, allergy relief, or sleep aid. To enhance validity of the sleep variables (Meltzer et al., 2012), sleep data (not cases) were excluded for participants who had fewer than five nights of sleep information. Participants with fewer than five nights of sleep data were more likely to be AA, p < .01 and to come from families with lower SES, p < .05. There were no other differences between those with and without complete sleep data for any demographic or main study variables.

Covariates

Parent report of family SES and adolescent sex were used as covariates in all analyses. Adolescents also completed the Revised Children’s Manifest Anxiety Scale (RCMAS; Reynolds & Richmond, 1978). This measure includes 28 items, with questions such as “I worry a lot of the time.” Responses are either “yes” (1) or “no” (0) and are summed to create a total anxiety score that was entered as a covariate in analyses. The scale had high reliability (α = .93).

Plan of Analysis

Two path models were fit, one for RSAB and one for RSAR. In each model, hypotheses were tested by examining community violence concerns, physiological regulation (RSAB or RSAR), race, and the two- and three-way interactions between them as statistical predictors of the sleep variables (minutes, efficiency), controlling for significant associations with the covariates.¹ Although significant two-way interactions are subsumed within three-way interactions, we interpreted both given the scant literature examining each moderator independently. This would make a meaningful contribution to the literature and facilitate comparisons between our findings and those of others who examine such questions in the future. Analyses were conducted using the AMOS Graphics add-on for SPSS 24, which allowed for simultaneous prediction of the two sleep outcomes. Furthermore, the covariances between the sleep variables were statistically controlled, which facilitated examination of unique associations between the predictor variables and each sleep outcome.

Per best practices, all variables were standardized to aid interpretation of effects. If analyses revealed a significant interaction effect, the variables were plotted at ± 1 SD and simple slopes were tested for whether they were significantly different from zero (Curran et al., 2004). Amos uses full information maximum likelihood estimation (FIML) to account for missing data, which results in the least biased estimates (Enders & Bandalos, 2001). Rates of missingness ranged from 5% (community violence concerns) to 27% (actigraphy variables), which is acceptable for use of FIML (Enders & Bandalos, 2001).

RESULTS

Preliminary Analyses

Bivariate correlations and descriptive statistics for the covariates and main study variables are included in Table 1. One outlier > 4 SDs from the mean (for community violence concerns) was winsorized to the next highest value. T-tests examining differences between EA and AA adolescents indicated that AAs had fewer sleep minutes than EAs (Table 2). Additionally, females had more sleep minutes (M_females= 413.11, SD=50.45; M_males=379.40, SD=61.84), t=3.79, p < .001, and greater sleep efficiency (M_females=92.52, SD=5.79; M_males=89.88, SD=8.09), t=2.31, p < .05, than males.

Table 1.

Descriptive Statistics and Correlations among Study Variables

	1	2	3	4	5	6	7	8	9
1. Socioeconomic status	–
2. Sex	.04	–
3. Anxiety symptoms	−.14	−.20**	–
4. Race	−.25**	−.09	.05	–
5. Community violence concerns	−.01	−.04	.47***	.11	–
6. Baseline RSA	.11	−.06	.09	.08	.09	–
7. RSA regulation	−.05	−.00	−.02	−.01	.01	−.28***	–
8. Sleep minutes	.10	−.29***	−.15	−.18*	−.07	−.06	.06	–
9. Sleep efficiency	.20*	−.19*	.00	−.11	−.01	.06	.11	.56***	–
Mean	2.90	–	10.01	–	2.25	6.84	−.46	398.05	91.34
(SD)	(1.83)	–	9.20	–	(4.17)	(1.23)	(.82)	(58.12)	(7.01)

Note. RSA = Respiratory sinus arrhythmia. Sex (0 = female; 1 = male); race (0 = White/European American; 1 = Black/African American); 398.05 minutes = 6.63 hours.

^*p < .05,

^**p < .01,

^***p < .001.

Table 2.

Descriptive Information by Race

	White/European American	Black/African American	t-test
	M (SD)	M (SD)
Community violence concerns	1.97 (3.67)	2.95 (5.14)	−1.35
Baseline RSA	6.77 (1.14)	6.99 (1.41)	−1.10
RSA regulation	−.46 (.86)	−.48 (.73)	.15
Sleep minutes	403.92 (58.91)	380.61 (52.62)	2.22*
Sleep efficiency	91.77 (7.18)	90.06 (7.18)	1.34

^*p < .05

Primary Analyses

RSAB and race as moderators of relations between community violence concerns and sleep

The first model examined community violence concerns, RSAB, race, the two-way and three-way interactions between them, and the covariates as predictors of the sleep variables (Table 3).

Table 3.

Unstandardized and Standardized Coefficients for Community Violence Concerns, Baseline RSA, and Race Predicting Sleep in Youth

	Sleep minutes			Sleep efficiency
	B (SE)	β	R2	B (SE)	β	R2
Socioeconomic status	–	–		–	–
Sex	−20.01***(4.25)	−.33		−1.41**(.52)	−.19
Anxiety symptoms	−12.03**(4.15)	−.20		–	–
Race	−13.15**(4.18)	−.22		−1.15*(.52)	−.15
Community violence concerns	−1.68(4.68)	−.03		−1.12*(.53)	−.15
Baseline RSA (RSAB)	−3.47(4.44)	−.06		.44(.55)	.06
			17%			7%
Comm. viol x RSAB	−4.13(5.08)	−.06		.84(.63)	.10
Comm. viol x race	−2.97(3.82)	−.06		−1.35**(.48)	−.20
RSAB x race	5.32(4.10)	.10		−.19(.51)	−.03
			19%			13%
Comm. viol x RSAB x race	9.46*(3.73)	.19		2.29***(.46)	.36
			24%			27%

Note. Path models controlled for sex, socioeconomic status, and anxiety symptoms when significantly associated with outcomes. Path coefficients reported are from the final model. R² reported is from the step of entry. SE = standard error. Sex (0 = female; 1 = male); race (0 = White/European American; 1 = Black/African American).

^*p < .05;

^**p < .01;

^***p < .001.

There were three significant direct effects. AA race was associated with fewer sleep minutes and poorer sleep efficiency, and greater community violence concerns were linked to poorer sleep efficiency (Table 3). There was one significant two-way interaction. Race moderated the association between community violence concerns and sleep efficiency (Figure 1). Among AAs, those with greater community violence concerns had poorer sleep quality than those with lower concerns. Furthermore, in the context of greater violence concerns, AA adolescents had poorer sleep efficiency than their EA counterparts, who had good sleep quality regardless of violence concerns.

Figure 1.

Two-way interactions between community violence concerns and race predicting sleep efficiency. AA = Black/African American; EA = White/European American; ns = nonsignificant.

There were two significant three-way interactions between community violence, RSAB, and race predicting sleep minutes (Figure 2a) and efficiency (Figure 2b) (Table 3). For AA adolescents with lower RSAB, a significant slope emerged in associations between community violence concerns and sleep efficiency. AA adolescents with greater community violence concerns and lower RSAB had lower sleep efficiency than AA youth with fewer concerns. All other adolescents had better sleep regardless of community violence concerns. There was a similar pattern of effects predicting sleep minutes, although none of the slopes reached significance.

Figure 2.

Three-way interactions between community violence concerns, baseline RSA, and race predicting (a) sleep minutes and (b) sleep efficiency. RSA = respiratory sinus arrhythmia. AA = African American; EA = European American; ns = nonsignificant.

RSAR and race as moderators of relations between community violence concerns and sleep

The second model examined community violence concerns, RSAR, race, the two-way and three-way interactions between them, and the covariates as predictors of the sleep variables (Table 4).

Table 4.

Unstandardized and Standardized Coefficients for Community Violence Concerns, RSA Regulation, and Race Predicting Sleep in Youth

	Sleep minutes			Sleep efficiency
	B (SE)	β	R2	B (SE)	β	R2
Socioeconomic status	–	–		1.15*(.51)	.15
Sex	−21.05***(4.23)	−.35		−1.60**(.51)	−.21
Anxiety symptoms	−12.58**(4.20)	−.21		–	–
Race	−13.54**(4.16)	−.22		−.86(.53)	−.11
Community violence concerns	−1.92(4.67)	−.03		−.94(.52)	−.12
RSA regulation (RSAR)	1.76(4.43)	.03		.35(.54)	.05
			18%			10%
Comm. viol x RSAR	−9.35*(4.57)	−.15		−2.07***(.55)	−.27
Comm. viol x race	−3.63(3.82)	−.07		−1.10*(.47)	−.16
RSAR x race	4.30(4.76)	.07		.20(.58)	.02
			21%			16%
Comm. viol x RSAR x race	−8.53*(3.63)	−.17		−2.18***(.44)	−.35
			26%			32%

Note. Path models controlled for sex, socioeconomic status, and anxiety symptoms when significantly associated with outcomes. Path coefficients reported are from the final model. R² reported is from the step of entry. SE = standard error. Sex (0 = female 1 = male); race (0 = White/European American; 1 = Black/African American).

^*p < .05;

^**p < .01;

^***p < .001.

Of the main study variables, one was directly associated with sleep. Specifically, AA race was associated with fewer sleep minutes. There were three significant two-way interactions. RSAR moderated the association between community violence concerns and both sleep variables (Figure 3a and 3b). The pattern of effects demonstrated that greater community violence concerns were associated with shorter sleep and lower sleep efficiency for adolescents who showed less RSA suppression in response to challenge (the less optimal response). There was also some indication that less RSA suppression in the context of fewer community violence concerns was protective. There was not a significant association between community violence concerns and either of the sleep variables for adolescents who showed greater RSA suppression. Additionally, there was a significant interaction between community violence concerns and race predicting sleep efficiency (Figure 3c). This interaction was identical to that detected in the first model and indicated that AA, but not EA, adolescents had poorer sleep efficiency in the context of greater community violence concerns.

Figure 3.

Two-way interactions between community violence concerns and RSA response to the stressor predicting (a) sleep minutes and (b) sleep efficiency; two-way interaction between community violence concerns and race predicting (c) sleep efficiency. RSA = respiratory sinus arrhythmia. AA = African American; EA = European American; ns = nonsignificant.

There were two significant three-way interactions between community violence, RSAR, and race predicting sleep minutes (Figure 4a) and efficiency (Figure 4b) (Table 4). For AA adolescents who showed less RSA suppression in response to challenge (less optimal response), significant slopes emerged in associations between community violence concerns for both sleep parameters. At higher levels of community violence concerns, AA adolescents who showed less RSA suppression experienced fewer sleep minutes and lower sleep efficiency. In contrast, AA adolescents who showed greater RSA suppression experienced higher sleep efficiency at greater community violence concerns. Furthermore, as demonstrated in Figure 4b, there was some evidence that less RSA suppression was protective for the sleep quality of AAs who reported fewer community violence concerns. There was not a significant relation between community violence concerns and the sleep variables for EAs, regardless of physiological regulation.

Figure 4.

Three-way interactions between community violence concerns, RSA response to the stressor, and race predicting (a) sleep minutes and (b) sleep efficiency. RSA = respiratory sinus arrhythmia. AA = African American; EA = European American; ns = nonsignificant.

DISCUSSION

The present study examined whether adolescents’ physiological regulation and race moderated the associations between community violence concerns and objective measures of sleep duration and quality. The findings generally supported our hypotheses and consistently demonstrated that youth at most risk for sleep problems were AAs with less optimal physiological regulation. There was not a relation between community violence concerns and the sleep variables for EA adolescents regardless of physiological regulation. These novel results contribute to the literature concerning the negative effects of community violence on sleep by identifying characteristics of adolescents who may be at greater risk for poor sleep in this context.

The findings are largely consistent with prior research demonstrating that lower RSAB and less RSA suppression increase risk for poor sleep (Elmore-Staton et al., 2012; El-Sheikh et al., 2013, separate dataset from current study; Irwin et al., 2006), particularly in the context of greater stress such as marital conflict (El-Sheikh et al., 2015, separate dataset from present study). The present study extends this work by demonstrating that physiological regulation may also moderate the relation between youth’s perceptions of distal environmental stressors and their sleep. Lower RSAB and less RSA suppression may indicate less adaptive regulation of arousal (Porges, 2007), which may be particularly important in contexts in which adolescents experience worries about their safety. Furthermore, the current study demonstrates that these relations exist among older adolescents in addition to the samples of younger children and preadolescents examined in prior research. Consistent with cumulative risk and health disparities work (Gee et al., 2012; Levy et al., 2016), these effects were found for AA and not EA adolescents. Of note, there was not a significant difference in the amount of community violence concerns reported by AA and EA adolescents in the present study. That vulnerability to poor sleep was largely specific to AA adolescents with poorer physiological regulation may suggest that race-based stressors place additional strain on the adolescent that when combined with difficulty regulating arousal makes it more difficult to sleep. In prior work, experiences of discrimination have been shown to disrupt both objective (Tomfohr et al., 2012) and subjective measures of sleep (Huynh & Gillen-O’Neel, 2016).

Coping strategies may be of particular importance for these adolescents, who experience higher levels of stress than their peers. Prior work has shown that among lower income, AA preadolescents, greater self-reported help-seeking from another person in response to general stress was protective against fewer sleep minutes and reduced sleep efficiency (El-Sheikh et al., 2014). Seeking support from others can be considered an adaptive, active coping strategy (Compas et al., 2001). As theorized indicators of an individual’s ability to appropriately direct physiological resources to environmental demands, higher RSAB and greater RSA suppression may underlie active coping. Improvement in behavioral coping as a result of intervention could translate to improved regulation of physiological arousal.

Some effects detected in the current study were unexpected. For example, the positive relation between community violence concerns and sleep efficiency for AA adolescents who showed greater suppression to the stressor was not hypothesized. Additionally, there was some indication that less RSA suppression in response to the stressor was associated with better sleep efficiency at lower levels of community violence concerns among AA adolescents. This result is contrary to literature suggesting that greater suppression is the most adaptive response. However, some recent work has noted that greater RSA suppression is more likely to be found as adaptive in samples with higher percentages of EAs (Graziano & Derefinko, 2013). It is possible that less suppression is advantageous in some contexts for populations that generally experience higher levels of stress. This result highlights the need to further examine adaptive patterns of physiological regulation among racially diverse samples.

There are several strengths and limitations of this study. Multiple objective sleep parameters were assessed, consistent with best practices in the field (Sadeh, 2015). Both sleep variables were included in the same models and were covaried, which allowed for testing of unique effects pertaining to each sleep parameter. However, actigraphy is not particularly well-suited to assess sleep latency (Martin & Hakim, 2011), and this may be an important sleep measure to consider because difficulty regulating arousal in stressful contexts is likely to manifest as longer sleep latency. Thus, we speculate that the lack of effects detected for sleep latency in the present study may be due to measurement, and that use of alternate tools, such as home polysomnography, may be fruitful.

AA and low SES adolescents were also less likely to have complete sleep data, although FIML was used to account for this missing information. The findings are specific to AA and EA adolescents living in rural and semi-rural communities in the U.S. who experienced relatively low levels of concerns on average, and may differ among adolescents who live in more urban locations, who experience more worry about violence in their community, and who are of different races. Whether similar processes are operative in the prediction of sleep among Hispanic or Latino individuals living in the U.S., who also experience race-based discrimination and disparities in sleep relative to EAs (Hale & Do, 2007), remains to be examined.

In sum, findings from the current study indicate that cumulative risk associated with greater concerns about community violence, less adaptive physiological regulation, and minority race is predictive of short and poor quality sleep. The results highlight the utility of studying multiple sleep parameters and identifying characteristics of adolescents who may be most at risk for sleep problems in the context of environmental stress. For some adolescents, psychoeducation concerning coping strategies for managing stress and arousal in stressful contexts may be beneficial for improving sleep duration and quality.

ABBREVIATIONS

EA

White/European American

AA

Black/African American

RSA

respiratory sinus arrhythmia

ECG

electrocardiograph

RSAB

Baseline respiratory sinus arrhythmia

RSAR

Respiratory sinus arrhythmia regulation

IGT

Iowa Gambling Task